A linear time approximation of Wasserstein distance with word embedding selection.
Sho OtaoMakoto YamadaPublished in: EMNLP (2023)
Keyphrases
- pointwise
- co occurrence
- distance measure
- distance function
- worst case
- error bounds
- approximation methods
- n gram
- approximation algorithms
- selection algorithm
- absolute error
- neural network
- euclidean distance
- string edit distance
- approximation error
- geodesic distance
- upper bound
- special case
- lower bound
- feature extraction